D&D General Explain Bounded Accuracy to Me (As if I Was Five)

[pemerton]

In 4e you never actually fought enemies wildly out of your level range. The instruction was to use roughly level appropriate opponents so that you would not never face those extreme numbers, that would break the game. The game even had awkward kludges like minions to deal with the math failure. 5e approach where the math just works and you can fight enemies of much lower and higher level without having to change the their stats is clearly superior.

I don't think it's clearly superior at all.

The 4e approach, with stable to-hit rates of around 50% to 70%, supports the system's approach to non-damage consequences of hits, both on the player-delivery side and the player-receiving side. It also supports more interesting play in general.

The 4e approach also makes a wider variety of fiction feasible: there is no incongruity, in 4e, in having a mid-paragon PC battle a Hobgoblin phalanx (in stat terms, a swarm), or in having a mid-epic PC battle a flight of Vrocks (again, in stat terms, a swarm).

Maybe they are 4e minions, or 4e swarms?

I would hate to be a player in a game where my DM makes this encounter in 5e; that's a LOT of dice rolling where the goblins need 17+ to hit only to do piddly damage. I suppose mathematically it's technically a threat, but I will be dead of boredom first, no?

Right. This is one of the things I have in mind - not the only one - in denying that the 5e approach is clearly superior.

As I posted upthread, the 5e approach is (on balance) less demanding on the GM. (I say "on balance" because the issue of numbers of opponents and action economy does seem to be a recurring point of discussion/concern in the context of 5e encounter building.)

As a 4e GM, I was never perturbed by the demands the system made on me.

 

[Ondath]

Yes and no. There is some great discussion in this reddit thread on it, but IMO there was a printed set of safety rails to help with learning them. The +4 attrib spells (bull strength/cats grace/owls wisdom/etc) were all 2nd level spells for both full casters (sorc/wiz cleric/druid/etc) and half casters (ie ranger/paladin). The full casters would unlock their first second level slot at level 3 & get their third 2nd level slot at level six, somewhere in that range it was common for players to get a +2 item for their prime attrib. Half casters by contrast got their first second level slot at level 10 when those X per caster level second level slots were really starting to make bull strength/casts grace/etc a hard sell, somewhere around here(give or take) it was common for players to pickup a +4 item for their base stat allowing half casters to drop an important +4 on secondary/tertiary attribs like a owls wis on the tank before the party faces something with a nasty wis save ability.

There was also the secondary guiderail of the not really named at the time iterative attack penalty that PF2 calls a multiattack penalty. Specifically attacks were made at +Full BaB/-5/-10/-15. The first one was very likely to succeed with the later ones less & less likely but still fairly reliable against mooks & such. Eventually monster ACs would scale to where the GM & players were likely to notice that mooks weren't very mooky & non-mooks were a bit too much missing; The GM would likely start dropping them or allowing players to buy/craft them by that point if they were previously uncertain.

Plus there was the wealth by level guidelines & I'd be shocked if the old dragon mag stuff didn't mention +attrib gear & +N magic items in the various letter to the editor/(ask [someone's name?]) section casually here & there if there wasn't some actual guidance at some point in one.

One of the major differences between then & now was that the presence of "Behind the Curtain" sections in the DMG & MM (probably other books too I bet) did a good job of talking about RAI and breaking down crunch for the reader to understand. That made it a lot easier for different GM's to be on the same page mentally when having a discussion.

Published adventures probably helped a lot too, +attrib gear was rather common for adventures of certain levels back then

edit: thinking about it there was a +2 attrib spell too I think, I don't remeber what they were but think it had some influence too

Um, I appreciate the comment but what you are explaining seems to be about 3.5. I was more interested in TSR-era D&D (so AD&D 2E and earlier), since I felt like the magic item curve for 3E onwards is fairly well explained.

 

[pemerton]

Um, I appreciate the comment but what you are explaining seems to be about 3.5. I was more interested in TSR-era D&D (so AD&D 2E and earlier), since I felt like the magic item curve for 3E onwards is fairly well explained.

I think it's fair to say that there is no clear expected magic item distribution for AD&D. It is possible to get some indications, from the magic item tables for NPC adventurers (in DMG Appendix C), from the magic item tables for preparing pre-gens/PCs above 1st level (in DMG Appendix P), and (as posted upthread) in the pre-gens found in some of the published modules.

But those three methods are not really consistent across themselves. For instance, the Appendix C tables give a 12th level character a chance of around 20% of having a weapon or armour of +3 or better. In Appendix P that chance is more like 10%. The pre-gens tend to be better equipped than either of the Appendices will allow for, without very lucky rolls.

 

[Crimson Longinus]

I don't think it's clearly superior at all.

The 4e approach, with stable to-hit rates of around 50% to 70%, supports the system's approach to non-damage consequences of hits, both on the player-delivery side and the player-receiving side. It also supports more interesting play in general.

The 4e approach also makes a wider variety of fiction feasible: there is no incongruity, in 4e, in having a mid-paragon PC battle a Hobgoblin phalanx (in stat terms, a swarm), or in having a mid-epic PC battle a flight of Vrocks (again, in stat terms, a swarm).

None of this relies on rapidly escalating stats. If the desired effect was that enemies beyond certain range become unbeatable, using such escalation would make sense. But wasn't the goal in 4e. Every time you want to use much more powerful or weaker foes you actually need to change their stats so that they can be fought. So escalating stats are just a nuisance, and not needed for anything 4e did. If you want, you can have swarms to represent monsters, enemies with interesting powers etc just fine even with bounded accuracy.

Now another matter is what you already acknowledged earlier: that objective rather than subjective stats are just easier for the GM to use. And it also creates clearer fiction to rules connection when same thing is not represented by wildly different mechanics at different times.

 

[Crimson Longinus]

Which is great for your group(s)!

Doesn't really affect whether there's a problem in general, or a disconnect between how the rules were designed to be used (in ways that make its structures effective/useful) and how the rules often are used at actual tables.

And as I've said before, I genuinely do not understand why this is a thing. The actual 5e text doesn't support doing it, and in some places even explicitly rejects it. It does not use a 3e-like skill system; if anything, the skills genuinely are closest to 4e (though without the explicit textual support for what I liked best about 4e's skill system proper, not counting stuff like SCs or the like, just the skills themselves.)

5e skills should be so much better than they are in practice, and I'm not alone in this feeling. Yet for some reason, the negative patterns of behavior that plagued 3e have been substantially inherited by 5e even when they shouldn't--even when it's less fun for everyone involved, including the GM! It baffles me to this day.


Well, I seem to be playing with a good 5e GM now.

I've had a couple duds with previous systems, including 4e, but never anything near as bad nor as frequent as with 5e. It's a bit part of why I'm such a skeptic about the whole "DM empowerment" crusade and the designers' tacit (or sometimes not-so-tacit...) "meh, you're the DM, you figure it out" attitude regarding the game's design.

The vast majority of 5e GMs I've had, I would rate no better than "medicore." Usually a lot lower. Even when said GM is a friend (which has, in fact, been the case two or three times.) With a system that puts GM skill, judgment, and responsiveness as the end-all, be-all of game design...well. I think you can see where that has tended to lead.

(Just in case you're reading, @Hussar, you are not in the above majority, but rather are the "good 5e GM now" I mentioned. Even though it hasn't been that long, I have already had a good time.)

As I've said before, 5e does crap job of actually teaching the GM to run the game. It tends to work well for experienced GMs who can just import their best practices from previous games. But for noobs there is way too much "do whatever, you figure it out." I'm sure Matt Mercer has done far more to teach new GMs how to run the game than the DMG ever has.

I think the skill DCs that you mention are good example of this. There really are no proper guidelines how to handle these, so people will end up assigning them in wildly different ways. I get that you cannot codify each instance, but a bunch of example would set some benchmarks and help to extrapolate consistently.

 

[MuhVerisimilitude]

How?

In every other edition the Ogre has 47 hit points, fights like a 5 HD monster, gets one attack a round, etc. etc. whether that Ogre is fighting a party of raw 1st-level types or a party of 15th-level bangers. Those stats do not change.

Because you are thinking at an individual monster level. That is: You match monsters with the PCs. This is why the first encounter of a level 1 party is an aboleth. Because you adjust the fiction to match the level of the player characters.

Extending this variability to monsters makes perfect sense. You can have stronger or weaker individual monsters of the same type. There's no reason why all goblins have the same strength, just like there is no reason why all human NPCs are equally strong.

Non-negotiable difference here IMO; in that for my part if something like this doesn't have an in-fiction explanation that's consistent with the rest of the fiction (even if it's weird) then it flat-out shouldn't be there.

PCs set the precedent for what makes adventurers tick mechanically. Last I checked, henches are usually hired as associate adventurers (as opposed to non-adventuring hirelings); which forces them to follow that precedent.

Nothing forces anything to follow anything.

Were we still in the 3e-4e era I'd sadly be inclined to agree with you. But guess what? We ain't in those days any more, and multi-level parties are now happily playable again.

I'll make a car analogy, because car analogies are great: Being able to support a mixed-level party is to me like having a car where you can configure the steering wheel to be inverted (turn left and the car goes right, turn right and it goes left). Sure it is technically a feature, but the feature is so incredibly niche I have never heard of anyone requesting it, and I cannot imagine any situation where it is actually useful to anyone.

Class levels are a fact of the fiction. If they weren't, PCs wouldn't have them either.

So classes can be part of the fiction if you make that explicit, but that has no bearing on whether or not player characters have them. I'm not seeing the connection here at all, frankly.

1e (and from what I can tell, also 0e and 2e) had in general a much flatter power curve than did 3e or 4e.

Multi-level parties were not only viable there, they were expected and assumed due to the variable advancement tables by class, level drains, individual xp, and a bunch of other factors. There is no reason whatsoever all those things can't be ported straight into 5e; and even if they're not, multi-level play is still viable there due to the flatter curve.

But note that modern design has done away with level drain and individual XP because they are unnecessary and complicated. Level drain is a nightmare because it requires you to lose features as you lose levels, so you need to track everything that you gain at every level so that you can undo it in case you lose the level.

It's also the height of un-fun mechanics.

Contrast this with 3e, where being off by even just a single level made a mighty big difference.

Was it really, though?

 

[EzekielRaiden]

As I've said before, 5e does crap job of actually teaching the GM to run the game. It tends to work well for experienced GMs who can just import their best practices from previous games. But for noobs there is way too much "do whatever, you figure it out." I'm sure Matt Mercer has done far more to teach new GMs how to run the game than the DMG ever has.

I think the skill DCs that you mention are good example of this. There really are no proper guidelines how to handle these, so people will end up assigning them in wildly different ways. I get that you cannot codify each instance, but a bunch of example would set some benchmarks and help to extrapolate consistently.

Of course, as soon as you start suggesting something in this direction, you are met with at least one of the following arguments:

1. "You can't protect players from bad GMs."
2. "No rule can be a perfect fit for every table."
3. "It's not the game's job to teach people how to run it."
4. "Anything you make will just be white-room theory."

#3 has become rather rarer in recent years, in large part due to the flip in consensus on the writing quality of the DMG. We've already seen #1 in this very thread, and I'm fairly sure someone's mentioned #2 at some point as well.

As I've said more than once, 5e overall has a culture of play antagonistic to the very idea of game design being a tool that can be made better. Pockets of resistance to that notion do exist, some of them focused on 3PP options. But by and large there is a pretty committed belief that, because the rules cannot be perfect, because they cannot prevent abuse, because they cannot guarantee improvement, because they aren't custom-tailored to every single individual case, there's nothing to be gained from even trying.

You already know how I feel about this stance. This has turned "the perfect is the enemy of the good" into a bitterly hilarious pantomime. Rather than bother even lifting a finger toward what benefits can be gained from such things, the whole project is completely written off because it isn't perfect. Meanwhile, the alternative is never perfect and if you challenge it on the exact same standard, suddenly expecting perfection is totally out of line.

 

[tetrasodium]

Because you are thinking at an individual monster level. That is: You match monsters with the PCs. This is why the first encounter of a level 1 party is an aboleth. Because you adjust the fiction to match the level of the player characters.

Extending this variability to monsters makes perfect sense. You can have stronger or weaker individual monsters of the same type. There's no reason why all goblins have the same strength, just like there is no reason why all human NPCs are equally strong.




Nothing forces anything to follow anything.



I'll make a car analogy, because car analogies are great: Being able to support a mixed-level party is to me like having a car where you can configure the steering wheel to be inverted (turn left and the car goes right, turn right and it goes left). Sure it is technically a feature, but the feature is so incredibly niche I have never heard of anyone requesting it, and I cannot imagine any situation where it is actually useful to anyone.

So classes can be part of the fiction if you make that explicit, but that has no bearing on whether or not player characters have them. I'm not seeing the connection here at all, frankly.

But note that modern design has done away with level drain and individual XP because they are unnecessary and complicated. Level drain is a nightmare because it requires you to lose features as you lose levels, so you need to track everything that you gain at every level so that you can undo it in case you lose the level.

It's also the height of un-fun mechanics.


Was it really, though?

Starting out with the tropical but, bounded accuracy prevents the (low)mixed levels in a mixed level party from feeling it & that is a terrible thing for everyone else at the table. I'll use an example from one of my campaigns where a player I'll call Alice would regularly show up late. The lateness started in the 5-10 or so minute range for reasonable reasons & eventually grew into an hour or more into a 2-3 hour session*. If Alice keeps up with the group then she is expected to be a full PC in encounters that the GM will either need to adjust or accept that it might be too tough for the party minus Alice. If Alice continually falls further & further behind as levels advance it doesn't have as much impact when she shows up midway into fight two or three of the session and opens with a fight ending nova starting with her biggest guns down the line knowing that the group needs to rest after needing to make up for the chilli's she's eating. Every time she drops a level behind the group it ensures that her absence will make less and less disruption to encounters she misses while the group is less & likely to be put out by the missing party member or find an interesting boss battle melted by fresh PC's reckless nova -- All of those combined make encounter creation easier for the GM to manage too.

There are lots of reasons that a player might fall behind though, choosing to come in with a new PC after making the GM kill the old one so they could distribute old magic items across the group and push for new ones a $level PC would have collected is a common one IME. By coming in at a lower level it puts some cost on the player doing it and minimizes the impact of "noob with a high level purchased character they don't know how to play in the raid"☆ type effects while they learn to play and decide how to grow the character just built so it fits the group. Skipping the grow to fit the group part often tends to result in a player making a PC to solve what they felt was a particular problem that was maybe only planned to be an issue for a couple sessions there is now a PC devoted to deleting. These too are good reasons why mixed level parties should be possible in a way that the low level PC feels it.


* For reasons like this

Spoiler

• Someone: Everyone remember that the game is at $ time today
• Alice-text:"I want to stop and order food since I've been at work all day"
• Over_optimistic_Group-text: Yea most of us too, that's why we started waiting a bit to order pizza so you could get here before it gets delivered or close enough that it's still hot.
• Alice-text: "oo that's right I'll think about it"
• Group stalling the GM's efforts to start: Anyone heard from alice? >someone pings Alice
• Alice-text: "I'm getting off off work now & clocking out in a minute, should be on my way soon"
• GM: "ok I see where this is going, lets start now"
• group: butbutbut Alice is out and we will be down a person, that's going to make things tough on us >fine lets order pizza
• 20minutes later.... anyone know where the heck Alice is? She only works like 2 miles away
• Alice-text: "I got off work a ago and just driving to $-local resteraunt to put in a food order because I've been at work all day and want to eat"
• Pizza arrives
• GM: we are starting now
• players butbutbut...
• 10-20 minutes later
  • GM: someone find out what alice is doing
• Alice-text: There was a wait at $applebees but I finally put in my order & just waiting on food now
• annoyed GM who saw this coming again when alice started today: Ok we are starting now
• Group: butbutbut
• GM30-40 minutes into a 2 hour session that hasn't even begun: no. Roll initiative >butbutbut>bob what's your AC
• Fight ends & group dithers a bit hoping to give Alice more time but really wants to check on Alice before they open the door & charge in or whatever
• 5 minutes later Alice-Text: There was some delays in the kitchen/ they got my order wrong & had to do it again. My food is coming out soon
• I wish this was hyperbole.

.

☆ wow & other MMOs must have coined a word for those by now but it's not coming to mind & my googlefu was weak this morning

 

[MwaO]

I think it's fair to say that there is no clear expected magic item distribution for AD&D. It is possible to get some indications, from the magic item tables for NPC adventurers (in DMG Appendix C), from the magic item tables for preparing pre-gens/PCs above 1st level (in DMG Appendix P), and (as posted upthread) in the pre-gens found in some of the published modules.

But those three methods are not really consistent across themselves. For instance, the Appendix C tables give a 12th level character a chance of around 20% of having a weapon or armour of +3 or better. In Appendix P that chance is more like 10%. The pre-gens tend to be better equipped than either of the Appendices will allow for, without very lucky rolls.

Or you could have played the adventures or looked at Rogue's Gallery that made it quite clear that Gygax was an extremely Monty Haul DM with PCs with crazy high stats who railed against those who were even more Monty Haul and pcs with higher stats than his table's PCs.

There's a Ring of Air Elemental Command in L1 as an example, one of the most "expensive" magic items in DMG, an adventure for level 2-4 PCs. Robilar, a Fighter, solo'd Temple of Elemental Evil with a Ring of Invisibility, Regeneration, a belt of giant strength, and a few other fantastic items.

 

[Charlaquin]

Of course, as soon as you start suggesting something in this direction, you are met with at least one of the following arguments:

1. "You can't protect players from bad GMs."
2. "No rule can be a perfect fit for every table."
3. "It's not the game's job to teach people how to run it."
4. "Anything you make will just be white-room theory."

#3 has become rather rarer in recent years, in large part due to the flip in consensus on the writing quality of the DMG. We've already seen #1 in this very thread, and I'm fairly sure someone's mentioned #2 at some point as well.

As I've said more than once, 5e overall has a culture of play antagonistic to the very idea of game design being a tool that can be made better. Pockets of resistance to that notion do exist, some of them focused on 3PP options. But by and large there is a pretty committed belief that, because the rules cannot be perfect, because they cannot prevent abuse, because they cannot guarantee improvement, because they aren't custom-tailored to every single individual case, there's nothing to be gained from even trying.

You already know how I feel about this stance. This has turned "the perfect is the enemy of the good" into a bitterly hilarious pantomime. Rather than bother even lifting a finger toward what benefits can be gained from such things, the whole project is completely written off because it isn't perfect. Meanwhile, the alternative is never perfect and if you challenge it on the exact same standard, suddenly expecting perfection is totally out of line.

It’s not “we can’t prevent abuse so we shouldn’t even try,” it’s “let’s not let the potential for abuse hamper our designs.” One of the weak points of 3e was that in trying to create a unified experience across tables, they ended up with an overwhelming mess of hard rules for hyper-specific cases. It’s to 5e’s credit that it moved away from this in favor of embracing flexibility and allowing different DMs to tailor their own games to their own preferences.

That doesn’t mean there’s no room for improvement in how the game teaches itself to new DMs. I’m very much looking forward to the revised DMG to help address this problem.